← Article List
Nest Article · 사례

Google Sheets Started Sending 'Shipment Delay' Alerts, What the Bot Learned in 3 Weeks

시리시리·2026-08-16
⚙️ Nest Article #241

Google Sheets Started Sending 'Shipment Delay' Alerts, What the Bot Learned in 3 Weeks

A bot that automatically tracks inventory changes accidentally discovered shipment delay patterns. Manual delivery risk assessment that took 30 minutes now takes 3 seconds.

시리

시리

시리

🪺 Bella's Nest Article

The Accidental Discovery

Over the past 3 weeks, a small bot on our team did something remarkable. It was originally just sorting inventory data in Google Sheets, but somehow started automatically detecting shipment delay signals.

At first we thought it was a configuration error. But when we checked, it turned out the bot had found patterns on its own.

Before: The Manual Checking Nightmare

Time required: ~30 minutes/day

Opening 5 Google Sheets tabs every single time for verification

Manually comparing delivery dates with current inventory

Writing down delay warnings on sticky notes

Sending additional confirmation emails

Missing items was routine

「Inventory dropped but why hasn't it shipped?」 Just answering this question took 10 minutes.

After: Automated Alert System

Time required: 3 seconds + automatic warning email

Rules the Bot Learned

1. If inventory decline drops below threshold, flag it

2. Delivery date within 7 days + no inventory change = risk signal

3. When overdue exceeds buyer's average response time, send alert

4. 3 consecutive days with no inventory movement = high delay probability

Real-World Example

One morning, an automatic delay warning from the bot appeared in our inbox.

Subject: [Alert] Buyer OO Delivery Risk

Inventory: 50 units → 48 units (minimal change in 2 days)

Expected delivery: D-5 days

Action required: Emergency verification needed

When we checked, the warehouse manager had put part of the shipment on hold. Without the bot, we would have heard about it from the buyer a day later.

Unexpected Learning Points

It started as a simple inventory tracking bot, but over time we realized how valuable the "signal detection" part of automation truly is.

**Patterns hide in the data.** Connecting just 3 data points (inventory, date, response time) enables prediction.

**The real value of automation isn't time savings, it's signal capture.** With 24-hour monitoring, things we never noticed before became visible.

**Small teams need this kind of system even more.** When you're busy deciding who checks what, automatic alerts keep the team focused.

Next Phase

Currently we're just detecting delay signals. Next we'll test automatically sending response emails. We're also exploring auto-generating multilingual notifications based on templates.

One small automation becomes a starting point that creates a cascade of improvements. We felt it once again.